Barry Marshall, MD, an Australian physician who won the Nobel Prize for discovering the link between stomach ulcers and the bacteria H. pylori, wants to change that. He’s devised what may be a simple, noninvasive way to diagnose IBS — by listening to the sounds produced in patients’ digestive tracts. He’s calling it the Noisy Guts Project. The Noisy Guts Project is testing the use of an acoustic belt that listens to and records the grumblings of your gut. The recordings are then fed into an AI (artificial intelligence) engine to see if it can detect the difference between IBS gut noise, and noise from healthy people. (People with IBS have stomach contractions, known as peristalsis, that are either too strong or that get out of sync, which creates pressure in different places in the gut.) If his diagnostic method works, it would be welcome news for millions of IBS patients around the world, who usually have to endure months or years of tests and questionnaires before a doctor can make a proper diagnosis. IBS is a common gastrointestinal disorder, affecting up to 15 percent of the population, according to the American College of Gastroenterology. It can produce diarrhea, constipation, and bowel incontinence. The disease disproportionately affects women, and often strikes young children. Marshall believes this leads many clinicians to make ungenerous assumptions about IBS patients. “Eighty percent of IBS patients are women, so doctors see it and they think ‘Oh, it’s stress.’ Same with little kids. They say, ‘It’s a school disease.’ My hope for this project is that it starts to show these doctors that it’s not just stress or anxiety, we’re dealing with a legitimate illness here,” Marshall says. But the Noisy Guts Project needs more data before it can offer a better way. “We’re in the proof-of-concept phase,” Marshall says. For now, results look promising. The project’s AI component can detect IBS at 87 percent sensitivity — meaning it can correctly identify IBS patients about 9 times out of 10. But this is not quite enough to satisfy gastroenterologists; they want to see more patients tested with the device before they trust it in their clinical practice. But every case makes the AI smarter. As time goes on, and the AI processes more patient data, it will become better trained to spot the characteristic noises of IBS. “All that’s left is to get enough data together to get regulators excited. Shouldn’t take more than a few years,” Marshall says. For now, Marshall and his team are developing a more sophisticated product that will filter out meaningless noise from the recordings. From there they hope to focus their research on gut bacteria to better understand its relationship with IBS. “The microbiome is a very confusing arena for IBS research,” he says. “The Noisy Guts Project can help steer it by identifying patients quickly, or by segmenting out IBS patients by what their symptoms are. Then we can take a closer look at the bacteria in their systems.”
